Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Sony A7S - Viable for EAA?

  • Please log in to reply
181 replies to this topic

#26 mclewis1

mclewis1

    Thread Killer

  • *****
  • Posts: 18665
  • Joined: 25 Feb 2006
  • Loc: New Brunswick, Canada

Posted 23 June 2015 - 07:57 PM

One of the really big issues with the evolution to larger sensors is as Don and others have mentioned ... the optics. For EAA/live viewing we are generally used to working at ratios like f3 to f5. Larger sensors start to put us in the same league optics expensive wise as the imaging folks who traditionally use reducers in the .8x range on more expensive scopes ... only we have an even more extreme problem.

 

Going down to the f5 and lower ranges with larger sensors is going to require field flatteners as well as reducers with enough aperture to keep vignetting at bay. That means a lot more money for the optics.

 

We either start with faster optics (f4 Newtonians) and add field flatteners (proven and generally reasonable price wise), or the extra complexity and expense of larger aperture and better corrected reducers on slower scopes (f7 refractors or f8 RCs for example).

 

It's going to become increasingly tough using our beloved native f10 SCTs with bigger sensors at the f ratios we've become accustomed to.



#27 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 23 June 2015 - 08:16 PM

My concern with the A7S is that the sensor's pixels are 8.4u x 8.4u and it doesn't appear that it's sensitivity is much different than ICX828/829 cameras.   If I were using the same scope at the same FR there would be no resolution gain compared to cameras like those using  ICX828/829 because of the similar pixel size.  Those cameras that use these sensors like the Xterminator, SR-DS, LSX2 all provide the same resolution for image size and are cheaper and easier to use and exposures are just as short or shorter. 

Maybe I'm missing something here.  If the A7S was more sensitive that these other cameras then one could use the same scope at a higher FR; thus longer FL, and end up with higher resolution for the same length of exposure as these other cameras. 

Where I think the A7S would shine would be in video mode if the frame rate could be changed from 1/4sec to something greater like 1sec or more.

 

The A7S has a full-size sensor. It's area is 25x larger than that of the ICX829. This makes an enormous difference and should not be overlooked.

 

Resolution is not identical with pixel size. Resolution = Number of pixels / Image area.
If one frames the same area of the sky with an ICX829 based camera and with a A7S, the latter would render the same objects with 25x more pixels and hence could resolve much finer detail. The assumption of using the same scope for both cameras is not justified. An object filling the entire field of the ICX829 would be tiny with the A7S, filling only 1/25th of the image area.

 

The large sensor, of course, puts high demands on the optics. To fully illuminate a full-size sensor with a flat field, one needs at least an 11"EdgeHD, a TV 127 IS or a Tak Epsilon. They cost $3000 and up.

 

NSN doesn't have the bandwidth to properly transmit and display images made with full-frame or even crop sensor cameras. It was conceived with the 0.5Mpx images of the old analog TV standards in mind. It fits the needs if the IXC829 but is not the right framework to assess higher quality imaging devices.

 

Clear Skies!
--Dom

 

Edit: I didn't see Mark's post while typing mine, hence the overlap.


Edited by Dom543, 23 June 2015 - 08:19 PM.

  • mclewis1 likes this

#28 DonBoy

DonBoy

    Apollo

  • *****
  • Posts: 1239
  • Joined: 06 Sep 2007

Posted 23 June 2015 - 08:46 PM

 

The A7S has a full-size sensor. It's area is 25x larger than that of the ICX829. This makes an enormous difference and should not be overlooked.

 

Resolution is not identical with pixel size. Resolution = Number of pixels / Image area.
If one frames the same area of the sky with an ICX829 based camera and with a A7S, the latter would render the same objects with 25x more pixels and hence could resolve much finer detail. The assumption of using the same scope for both cameras is not justified. An object filling the entire field of the ICX829 would be tiny with the A7S, filling only 1/25th of the image area.

 

Dom,

 

25X the resolution with the same frame size with the A7S would require a scope 5 times the FL for the same FR and yes result is 5x more arcsec/pixel.  But using the A7s on the same imagine with the same scope and same FR will have the same arcsec/pixel for either the A7S or an ICX829 no more resolution. 

 

Example using a C8 at f5 would yield a 1.73 arcsec/pixel image scale for the A7S and also the same for the ICX829.

 

If one sets the A7s to the same frame size as the ICX829 the scope would need to have a 40" aperture at f5 at a FL of 1000 vs 40" @f5 at a FL of 5000 and yes 5 times more arcsec/pixel.


Edited by DonBoy, 23 June 2015 - 08:48 PM.


#29 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10233
  • Joined: 03 Jul 2004

Posted 23 June 2015 - 09:51 PM

On "why A7S" or "why A7s has challenges", best way is to pull back to 100,000 ft to view the landscape.

 

From that perspective, one should notice in civilian / commercial mass-produced market, low-light and specialty market has never been the main focal point for these camera mfgs.

 

Occasionally, these camera firms will try something to test the water, but never as serious as a dedicated astro-imager / astro-camera vendor such as SBIG, QSI, Atik, QHY, etc.

Bear in mind, they really need volume or large enough profit to sustain their companies (including many talented engineers.)

Think: we are lucky enough at least there are some models to choose from.

 

Producing Full-Frame formatted camera makes financial sense (justify the body's high price tag) together with many premium lenses (image circle large enough for FF, fast f-ratio, etc.) purchases.   The latter is the biggest factor here.

 

Considering an A7s' price tag is $2500 which is a full-frame sized mirrorless.  Comparing that to an image-head (no DSP, no LCD, no API to control, need a PC to make it to work), say, using a type-1/1.5" (which is type-2/3") such as ICX825, the sensor size difference is 1:0.067 (about 14.9 times larger, area wise) typically cost $1200, or

comparing that with a type-1/2" imager such as ICX829, the sensor size difference is 1:0.0361 (about 27.7 times larger) typically costs $600,

you would appreciate the mass-production driven economy benefit.

 

Next post: tech info

 

BTW, sensor sizes (diagram and table) here

http://www.cloudynig...e-sensor-sizes/

 

Clear Skies!

 

ccs_hello



#30 ccs_hello

ccs_hello

    Voyager 1

  • *****
  • Posts: 10233
  • Joined: 03 Jul 2004

Posted 23 June 2015 - 10:10 PM

In the low light image sensor, pixel pitch matters (more photons to be collected on a larger per-pixel area),

Bayer filter type (RGBG, CMYG, clear-RGB, etc.) matters,

sensor technologies used (e.g., BSI, coppoer in metal layer, doping, etc., many to be mentioned here) matters,

and

- keep the sensor quality high when high-gain is applied,  <-- fr this you need a sensor that is of high S/N design is the first place.

- high quality and high precision built-in A/D converter matters.

 

The above just address one single pixel, as if they are independent.  This is per-pixel quality metrics.

 

Beyond that, the overall image quality matters.  An image is formed by many discrete pixels.

More pixels in an image, i.e., higher spatial resolution, gives a better result.  Some call it "picture S/N" or "image quality".

With the equal amount of the per-pixel S/N, higher resolution image looks far better than low resolution image.

This is how human vision perceives the image. 

 

Back to A7s, my view is

I wish SONY will make a sensor that is just APS-C sized or even Four-Thirds sized,

since we the near-realtime folks just want to get the native pixels' output to a monitor (say a 4K monitor: 3840x2160 or 4096x2160.)

Smaller imager size lessens the burdens on fast-optics as well.  It's a win-win.  <-- I hope the trend goes this way soon.

 

Clear Skies!

 

ccs_hello



#31 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 23 June 2015 - 10:10 PM

Don,

 

I don't disagree with your numbers. But I also said this:

The assumption of using the same scope for both cameras is not justified. An object filling the entire field of the ICX829 would be tiny with the A7S, filling only 1/25th of the image area.

To put the other way around, this could also be formulated as follows. "If one insists on using the same scope at the same f-ratio, then the A7S cannot be justified". There are friends on this forum, who make excellent wide field images with their ICX829 camera attached to a 50mm finderscope. In terms of this example, what I am saying is that it is not justifiable to use the same finderscope with the A7S.

 

As also said by Mark and others, to match the potential of the A7S, we need more expensive optics (larger apertures) and/or need to move to higher f-ratios. Higher f-ratio will require longer exposures, a move away from EAA, or higher ISO settings. The beauty of the A7S is that it does allow very high ISO. This is what makes it unique and intriguing for us in EAA !

 

Clear Skies!

--Dom


Edited by Dom543, 24 June 2015 - 12:00 AM.


#32 Don Rudny

Don Rudny

    Surveyor 1

  • *****
  • Posts: 1622
  • Joined: 21 Mar 2013
  • Loc: Mauna Kea, Hawaii

Posted 23 June 2015 - 10:33 PM

Dom,

 

Did you mean higher f ratio?  If I used the same scope setup for my M57 with the A7s and wanted the same image scale, I would need a 4x Barlow that would make my F ration nearly 40 instead of 10 with my LS.  I think my exposure would need to be 16 times longer.  Am I correct on the math?


  • djgilley likes this

#33 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 23 June 2015 - 11:13 PM

Dom,

 

Did you mean higher f ratio?  If I used the same scope setup for my M57 with the A7s and wanted the same image scale, I would need a 4x Barlow that would make my F ration nearly 40 instead of 10 with my LS.  I think my exposure would need to be 16 times longer.  Am I correct on the math?

 

As in another thread, we should remember that is physics, not math, that applies here. The rule of thumb that is the 'math' you mention about f/ratio and exposure time may or may not apply across such different cameras, a lot depends on the physics of the actual light flux & pixels. Also, as mentioned, we can increase ISO as required. We don't even know yet how the ISO range of the A7s compares to those other cameras in common EAA use.  Lastly, the used prices for A7s have been dropping, and will probably continue to do so once the A7sII is out.



#34 Don Rudny

Don Rudny

    Surveyor 1

  • *****
  • Posts: 1622
  • Joined: 21 Mar 2013
  • Loc: Mauna Kea, Hawaii

Posted 24 June 2015 - 12:05 AM

Curtis,

 

I think math always applies as well as physics.  If the A7s is some magical camera that has more sensitivity than anything else, then, yes that factors in.  But, as pointed out in other threads there is no magic, so I don't think that its sensor will produce sensitivity or QE greater than what is available in the EAA sensors typically used now.  Increased gain and ISO produce a lot of noise and at some point becomes objectionable to the most casual observer.  



#35 schwim

schwim

    Viking 1

  • *****
  • topic starter
  • Posts: 598
  • Joined: 23 Oct 2014

Posted 24 June 2015 - 12:20 AM

Curtis,

 

Increased gain and ISO produce a lot of noise and at some point becomes objectionable to the most casual observer.  

 

But some things simply look better than others. At some point subjectivity rules, no matter the math or physics.

 

Really, what many of us want is the experience of seeing what we imagine. The question at hand is the A7S a top N contender as a tool for the job.



#36 Don Rudny

Don Rudny

    Surveyor 1

  • *****
  • Posts: 1622
  • Joined: 21 Mar 2013
  • Loc: Mauna Kea, Hawaii

Posted 24 June 2015 - 12:39 AM

 

Curtis,

 

Increased gain and ISO produce a lot of noise and at some point becomes objectionable to the most casual observer.  

 

But some things simply look better than others. At some point subjectivity rules, no matter the math or physics.

 

Really, what many of us want is the experience of seeing what we imagine. The question at hand is the A7S a top N contender as a tool for the job.

 

The job at hand is EAA, not imaging, so acquisition time is important, thus the math and physics.  Is the A7s a top contender for EAA?  I don't see it, but I could be wrong. 



#37 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 24 June 2015 - 12:48 AM

Thank you Don!

 

I should have said higher (worse) f-ratio. I edited my post to correct this.

 

Just as in the case of your namesake Don(Boy), I agree with your numbers. And given your choice of the object of the Ring nebula, I wouldn't contest your conclusion either. I don't see any reason to justify the use of the A7S over the Lodestar to capture M57.

 

But this doesn't mean that I am throwing in the towel. This past winter and spring I made a couple of Lodestar captures with my 85mm SMC Takumar portrait lens. Here are some examples.

http://stargazerslou...ions/?p=2657517
http://stargazerslou...lter/?p=2557186
http://stargazerslou...lter/?p=2559012

 

I would absolutely love to redo all those images with a Sony A7S using my 400mm f/2.8 Nikkor lens. No need to increase exposure times, both lenses are entirely real and, as much as I like my Lodestar, I am afraid that the A7S shots would be a class above. Consistent with the 25x pixel count and the 10x price of the Nikkor. (I mean this, not just saying it to confound those, who label me a rabid Lodestar fanatic.)

 

To make the long story short, for some extended objects, the A7s's large sensor would win out. For others, the Lodestar is hard to beat. If someone gave me a Sony, I would still keep and use my Lodestar.

 

Clear Skies!

--Dom


Edited by Dom543, 24 June 2015 - 01:02 AM.


#38 Don Rudny

Don Rudny

    Surveyor 1

  • *****
  • Posts: 1622
  • Joined: 21 Mar 2013
  • Loc: Mauna Kea, Hawaii

Posted 24 June 2015 - 01:08 AM

Yes, Dom, I agree.  There is no question that the large sensor will achieve it's full resolution with wide field objects.  I can't even get the NA nebula in my 162mm FL scope with the Lodestar.  But that's a niche area of EAA.  Overall, most desirable objects are in the 400 to 900mm FL with the 1/2" sensor.  Even on the wide field shots your Lodestar images were impressive.  In the world of EAA, it's tough to compete with the high sensitivity 1/2" sensor cams like the Mallicams and Lodestars.  Price is nice, too.



#39 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 24 June 2015 - 10:22 AM

To clarify, the focal length vs image time is a 'rule of thumb' from the film days. We are using vastly different systems at much lower light levels than that 'math' was created for. I'm a Physicist and to me critical thinking is important and hence why I question the applicability of the f-ratio vs image time 'rule', 'math' only applies to physical systems when it's proven to do so. I think it's enough to say that some testing is in order.



#40 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 24 June 2015 - 10:58 AM

I may sound as a broken record to say that we would like to see more images posted. Accompanied with a few basic numbers, like camera used, exposure time, gain, f-ratio, aperture and stacking info.

 

Then those, who like math could check if their formulas apply. And those, who prefer to learn directly from examples, bypassing math, could do that.

 

Just as an example, we have had a six page long thread about the ATIK414ex. I learned more from the few captures posted by DonBoy and Lyle in the gallery than from the 150+ posts of the discussion thread. I am not saying that discussion is not useful. Just that we would also like to see more real captures.

 

Clear Skies!

--Dom


  • mclewis1 likes this

#41 CHAPSKINS

CHAPSKINS

    Ranger 4

  • -----
  • Posts: 344
  • Joined: 27 Nov 2013

Posted 24 June 2015 - 11:37 AM

It's *maths* not "math"  :gaah: ...pfft...Mericans :)

 

The biggest stumbling block for the A7s is the software for longer live video times. Like I mentioned previously, this unfortunately doesn't look like it's going to change anytime soon in that regard as Sony have locked down their software with the exception of iApps...blah. 

 

I really wish that Sony would allow developers the opportunity to lengthen video times to more than the poultry 1/4 sec :(

 

I'm hoping the sky's will clear for the weekend and see what this camera can do. Something I've longed for is the wider views via my HyperStar, something my Mallincam X2 sucks at as good as it is and what it does. 

 

I should mention: I didn't pay the full list price for my A7s which might be of interest to others thinking about picking one up. I went through a company called Panamoz over here in the U.K that grey import them and offer a full guarantee, which when reading other people have bought from them, they do fully honour that guarantee. The Sony list price over here in the U.K is just over £2k, I payed just over £1.3k for it. If there's other companies that offer the same as Panamoz in your country, it's well thinking about so as to save a pile of cash :)


Edited by CHAPSKINS, 24 June 2015 - 11:39 AM.


#42 chasing photons

chasing photons

    Mariner 2

  • -----
  • Posts: 287
  • Joined: 19 Mar 2013
  • Loc: Minnesota

Posted 24 June 2015 - 11:44 AM

All of this math talk makes me want to go to hospital…  I mean THE hospital.   ;)



#43 CHAPSKINS

CHAPSKINS

    Ranger 4

  • -----
  • Posts: 344
  • Joined: 27 Nov 2013

Posted 24 June 2015 - 11:52 AM

I just stare at the maths lark like a dog being shown a card trick...

 

.



#44 Dragon Man

Dragon Man

    Gemini

  • *****
  • Posts: 3381
  • Joined: 07 Mar 2006
  • Loc: Snake Valley, Australia

Posted 25 June 2015 - 10:25 AM

It's *maths* not "math"  :gaah: ...pfft...Mericans :)

It's 'riffma tick' :gaah: ... pfft .... Murkins and Pommy Gits   :lol:


  • dvb likes this

#45 Dom543

Dom543

    Apollo

  • *****
  • Posts: 1099
  • Joined: 24 Oct 2011

Posted 25 June 2015 - 10:37 AM

This thread starts to sound as a first talking Barbie Doll convention.

The first sentence she uttered was "Math is tough".

It was quickly changed in response to an uproar by women's organizations and educators.

 

Let's get back to the OT.

--Dom



#46 schwim

schwim

    Viking 1

  • *****
  • topic starter
  • Posts: 598
  • Joined: 23 Oct 2014

Posted 25 June 2015 - 10:59 AM

 

Let's get back to the OT.

--Dom

 

Works for me.

 

As far as controlling the camera, I posted some information I found about Sony's APIs here. From reading the API documents, it seems pretty clear that its feasible to do most of the things we might want via software running on a PC or laptop. This got me excited, and I found an A7S at the local electronics store just so I could test this out. I'm *not* a programmer, but I can do some basic scripting.

 

In summary, I've been able to connect to the camera and interact with it at a very basic level. I haven't been able to get it to do anything yet, but I chalk that up to my lack of experience with software development than the camera's inability to be remotely operated by my PC.



#47 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 25 June 2015 - 12:37 PM

You can use RCC, only thing really missing is good bulb controls.

#48 schwim

schwim

    Viking 1

  • *****
  • topic starter
  • Posts: 598
  • Joined: 23 Oct 2014

Posted 25 June 2015 - 12:47 PM

You can use RCC, only thing really missing is good bulb controls.

 

True. I haven't tried it yet. Can it be automated?

 

EDIT: To clarify what I'm getting at, my EAA experience has been with Miloslick. Given what I see in the Sony API, it should be feasible to build a tool that does something similar. Alternately, a tool to just automatically trigger the camera as one sees fit while Astrotoaster/DSS does its thing might be good enough. I admit I don't know if DSS can do this, even within the range of non-bulb exposures.


Edited by schwim, 25 June 2015 - 12:55 PM.


#49 Relativist

Relativist

    Cosmos

  • *****
  • Posts: 8151
  • Joined: 11 Oct 2003
  • Loc: OC, CA, USA

Posted 25 June 2015 - 05:12 PM

In terms of automated, you can tell it to take a certain number of exposures, 1, 5 , 10, 53, whatever you type in. The intervalometer works for exposures 30 sec or less, and you can stop it if I remember correctly. Like I had said the only thing missing is variable exposures over 30s with intervalometer controls.

#50 CHAPSKINS

CHAPSKINS

    Ranger 4

  • -----
  • Posts: 344
  • Joined: 27 Nov 2013

Posted 26 June 2015 - 11:47 AM

 

 

Let's get back to the OT.

--Dom

 

Works for me.

 

As far as controlling the camera, I posted some information I found about Sony's APIs here. From reading the API documents, it seems pretty clear that its feasible to do most of the things we might want via software running on a PC or laptop. This got me excited, and I found an A7S at the local electronics store just so I could test this out. I'm *not* a programmer, but I can do some basic scripting.

 

In summary, I've been able to connect to the camera and interact with it at a very basic level. I haven't been able to get it to do anything yet, but I chalk that up to my lack of experience with software development than the camera's inability to be remotely operated by my PC.

 

 

Via the API, can you control the A7s via USB on your desktop computer?




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics